Hindawi BioMed Research International Volume 2021, Article ID 9921012, 17 pages https://doi.org/10.1155/2021/9921012 Research Article Identification of Macrophage Polarization-Related Genes as Biomarkers of Chronic Obstructive Pulmonary Disease Based on Bioinformatics Analyses Yalin Zhao , Meihua Li, Yanxia Yang, Tao Wu, Qingyuan Huang, Qinghua Wu, and Chaofeng Ren Respiratory and Critical Care Medicine, Kunming First People’s Hospital, Kunming, Yunnan Province, China Correspondence should be addressed to Yalin Zhao; [email protected] and Chaofeng Ren; [email protected] Received 31 March 2021; Accepted 4 June 2021; Published 21 June 2021 Academic Editor: Nagarajan Raju Copyright © 2021 Yalin Zhao et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objectives. Chronic obstructive pulmonary disease (COPD) is characterized by lung inflammation and remodeling. Macrophage polarization is associated with inflammation and tissue remodeling, as well as immunity. Therefore, this study attempts to investigate the diagnostic value and regulatory mechanism of macrophage polarization-related genes for COPD by bioinformatics analysis and to provide a new theoretical basis for experimental research. Methods. The raw gene expression profile dataset (GSE124180) was collected from the Gene Expression Omnibus (GEO) database. Next, a weighted gene coexpression network analysis (WGCNA) was conducted to screen macrophage polarization-related genes. The differentially expressed genes (DEGs) between the COPD and normal samples were generated using DESeq2 v3.11 and overlapped with the macrophage polarization-related genes. Moreover, functional annotations of overlapped genes were conducted by Database for Annotation, Visualization and Integrated Discovery (DAVID) Bioinformatics Resource. The immune-related genes were selected, and their correlation with the differential immune cells was analyzed by Pearson. Finally, receiver operating characteristic (ROC) curves were used to verify the diagnostic value of genes. Results. A total of 4922 coexpressed genes related to macrophage polarization were overlapped with the 203 DEGs between the COPD and normal samples, obtaining 25 genes related to COPD and macrophage polarization. GEM, S100B, and GZMA of them participated in the immune response, which were considered the candidate biomarkers. GEM and S100B were significantly correlated with marker genes of B cells which had a significant difference between the COPD and normal samples. Moreover, GEM was highly associated with the genes in the PI3K/Akt/GSK3β signaling pathway, regulation of actin cytoskeleton, and calcium signaling pathway based on a Pearson correlation analysis of the candidate genes and the genes in the B cell receptor signaling pathway. PPI network analysis also indicated that GEM might participate in the regulation of the PI3K/Akt/GSK3β signaling pathway. The ROC curve showed that GEM possessed an excellent accuracy in distinguishing COPD from normal samples. Conclusions. The data provide a transcriptome-based evidence that GEM is related to COPD and macrophage polarization likely contributes to COPD diagnosis. At the same time, it is hoped that in-depth functional mining can provide new ideas for exploring the COPD pathogenesis. 1. Introduction ity rates of COPD are still growing [4]. Worse, although COPD has long been considered treatable, the diagnosis Chronic obstructive pulmonary disease (COPD), an inflam- and treatment of COPD in the clinical practice remain to matory disease of the lung mainly caused by smoking tobacco be improved [5, 6]. Currently, the diagnosis of COPD mainly cigarettes and environmental exposure from burning bio- depends on the use of spirometry by identifying pulmonary mass fuels or air pollution [1, 2], has become a major health dysfunction [7], which is full of limitations, such as the detec- problem around the world [3]. It has been reported that tion of the early stage of COPD. Moreover, although some COPD killed about 3 million people in 2016, and the mortal- blood-related biomarkers were related to exacerbations, 2 BioMed Research International progression, or mortality of COPD, it was unknown whether stroke [22] and schizophrenia [23]. Moreover, all of these they could be selected as the diagnostic biomarkers [8]. On researches suggested that screening biomarkers for diagnosis the other hand, even if the traditional treatments including and treatment of complex disease by WGCNA was effective. lung volume reduction surgery for COPD have made great In the present study, we firstly downloaded the raw gene strides, there were still some uncontrollable complications expression profile dataset which included COPD samples [9]. In addition, even if some pulmonary rehabilitation pro- and normal samples from the Gene Expression Omnibus grams, such as exercise training, education, and behavior (GEO) database (https://www.ncbi.nlm.nih.gov/geo/query/ change followed by patient-tailored therapies have been acc.cgi). Next, WGCNA was carried out to screen macro- regarded as promising measures for improving the COPD phage polarization-related genes. Finally, we identified hub patients, a few patients still cannot gain the benefit from the genes as the novel diagnostic biomarkers for COPD and fur- exercise training [10]. Hence, it is very wishful for further ther analyzed their molecular mechanisms in COPD, which parsing the molecular mechanism and recognizing the novel will contribute to the treatment of COPD. biomarkers of COPD. With the development of bioinformatics, recent studies 2. Materials and Methods have revealed that COPD susceptibility is associated with the genes’ expression [11–14]. For example, it has been sug- 2.1. Dataset Acquisition. The raw gene expression profile gested that the expression of B3GNT, LAF4, and ARHGEF10 dataset (GSE124180, only peripheral blood samples) was can predict the frequent exacerbations of COPD [11]. More- collected from the GEO database [24]. The dataset based on over, IL6 and SOCS3 also have been suggested to play a key GPL16791 platform included 6 COPD samples and 15 role in COPD and can be used as the therapeutic targets of normal samples. COPD [12]. Moreover, Zhang et al. also found that TLR2 and CD79A may be used as the potential biomarkers for 2.2. WGCNA. The weighted gene coexpression network the clinical severity of COPD and related to the inflammatory analysis was performed using WGCNA R package (v1.69) responses of COPD [13]. More importantly, it has been in the COPD and normal samples [25]. A total of 35 revealed that COPD can be subdivided into three molecular macrophage polarization genes (Supplementary Table 1) subtypes based on the gene expression profiles of 213 from the MsigDB database (https://www.gsea-msigdb.org/ COPD-related genes [14]. Therefore, bioinformatics analysis gsea/msigdb) were considered different traits to investigate of gene expression profiling may contribute to screening the coexpressed genes related to macrophage polarization novel biomarkers of COPD resulting in improving the treat- genes. Soft thresholding was then applied by raising ment of COPD. correlation values to a power of 14 to amplify disparity Increasing evidences have proposed that macrophage between strong and weak correlations. The soft thresholding which is an important effector cell for the innate immune power was chosen to achieve approximately scale-free response plays a crucial role in COPD [15–17]. A recent network topology, as recommended for biological networks study has demonstrated that COPD patients exhibit more [26, 27]. The resulting signed adjacency matrix was used to macrophages in the bronchial alveolar lavage fluid, along calculate topological overlap matrix (TO), which was then the airways and lung parenchyma compared to the normal hierarchically clustered with (1-TO) as a distance measure. samples [15, 16]. Besides, Traves et al. also found more blood Genes were then assigned into coexpression modules by monocyte-derived macrophages into the airspaces of COPD dynamic tree cutting algorithm requiring minimal module size [17]. More importantly, emerging evidences showed that of 100 genes [28]. The modules with highly correlated eigen- macrophage polarization may be associated with COPD genes (correlation above 0.6) were merged. Module eigengene [18–20]. For example, a previous study has suggested an (ME) is the first principal component of the gene expression increase in proinflammatory M1 macrophages in the small values within a module and is used to summarize the module’s airways of COPD compared to controls, but a reciprocal expression. The Pearson correlation between each gene and decrease in M2 macrophages [18]. On the other hand, the ME was then calculated. This value is called module member- results of Eapen et al. suggested that lncRNA MIR155HG ship (MM) and represents how close a particular gene is to a modulated the progression of COPD by inducing the module. Finally, each gene was assigned to a module for which granulocyte-macrophage colony-stimulating factor-mediated it had the highest MM. The module with the highest absolute M1/M2 macrophage polarization [19]. In addition, Shaykhiev value of the correlation coefficient with the traits was chosen as et al. suggested that the reprogramming for alveolar macro- the key module for subsequent analysis. phage polarization likely contributes to COPD pathogenesis